glibc | GNU Libc - Extremely old repo

You should try to export the model using torch.onnx. The page gives you an example that you can start with.

An alternative is to use TorchScript, but that requires torch libraries.

Both of these can be run without python. You can load torchscript in a C++ application https://pytorch.org/tutorials/advanced/cpp_export.html

ONNX is much more portable and you can use in languages such as C#, Java, or Javascript https://onnxruntime.ai/ (even on the browser)

Just modifying a little your example to go over the errors I found

Notice that via tracing any if/elif/else, for, while will be unrolled

First of all, the state of the OpenMP task scheduling on NUMA systems is far from being great in practice. It has been the subject of many research project in the past and they is still ongoing project working on it. Some research runtimes consider the affinity hint properly and schedule the tasks regarding the NUMA node of the in/out/inout dependencies. However, AFAIK mainstream runtimes does not do much to schedule tasks well on NUMA systems, especially if you create all the tasks from a unique NUMA node. Indeed, AFAIK GOMP (GCC) just ignore this and actually exhibit a behavior that make it inefficient on NUMA systems (eg. the creation of the tasks is temporary stopped when there are too many of them and tasks are executed on all NUMA nodes disregarding the source/target NUMA node). IOMP (Clang/ICC) takes into account locality but AFAIK in your case, the scheduling should not be great. The affinity hint for tasks is not available upstream yet. Thus, GOMP and IOMP will clearly not behave well in your case as tasks of different steps will be often distributed in a way that produces many remote NUMA node accesses that are known to be inefficient. In fact, this is critical in your case as stencils are generally memory bound.

If you work with IOMP, be aware that its task scheduler tends to execute tasks on the same NUMA node where they are created. Thus, a good solution is to create the tasks in parallel. The tasks can be created in many threads bound to NUMA nodes. The scheduler will first try to execute the tasks on the same threads. Workers on the same NUMA node will try to steal tasks of the threads in the same NUMA node, and if there not enough tasks, then from any threads. While this work stealing strategy works relatively well in practice, there is a huge catch: tasks of different parent tasks cannot share dependencies. This limitation of the current OpenMP specification is a big issue for stencil codes (at least the ones that creates tasks working on different time steps). An alternative solution is to create tasks with dependencies from one thread and create smaller tasks from these tasks but due to the often bad scheduling of the big tasks, this approach is generally inefficient in practice on NUMA systems. In practice, on mainstream runtimes, the basic statically-scheduled loops behave relatively well on NUMA system for stencil although it is clearly sub-optimal for large stencils. This is sad and I hope this situation will improve in the current decade.

Be aware that data initialization matters a lot on NUMA systems as many platform actually allocate pages on the NUMA node performing the first touch. Thus the initialization has to be parallel (otherwise all the pages could be located on the same NUMA node causing a saturation of this node during stencil steps). The default policy is not the same on all platforms and some can move pages between NUMA nodes regarding their use. You can tweak the behavior with numactl. You can also fetch very useful information from the hw-loc tool. I strongly advise you to manually the location of all OpenMP threads using OMP_PROC_BIND=True and OMP_PLACES="{0},{1},...,{n}" where the OMP_PLACES string set can be generated from hw-loc regarding the actual platform.

For more information you can read this research paper (disclaimer: I am one of the authors). You can certainly find other similar research paper on the IWOMP conference and the Super-Computing conference too. You could try to use research runtime though most of them are not designed to be used in production (eg. KOMP which is not actively developed anymore, StarPU which mainly focus on GPUs and optimizing the critical path, OmpSS which is not fully compatible with OpenMP but try to extend it, PaRSEC which is mainly designed for linear algebra applications).

Perhaps my failed attempts will help someone else solve this problem; my approach:

• boot up a clean CentOS 7 vm
• install R and some dependencies

I would do it like that.

I had the same issues, and I downgraded to the following version:

According to man7.org getifaddrs, any of the socket operations could be a cause for EBADF

ERRORS

getifaddrs() may fail and set errno for any of the errors specified for socket(2), bind(2), getsockname(2), recvmsg(2), sendto(2), malloc(3), or realloc(3).

Unrelated, but do you do freeifaddrs() somewhere?

The fs segment register is used in x86-64 Linux to point to thread-local storage. See How are the fs/gs registers used in Linux AMD64? So this instruction will xor the rdx register with the value found at offset 0x30 in the thread-local storage block.

This code is part of a pointer encryption mechanism in glibc to help harden against certain exploits. There is some explanation of it at https://sourceware.org/glibc/wiki/PointerEncryption. The value at fs:0x30 is an "key" for a trivial "encryption" algorithm; pointers are xor'ed with this value (and then rotated) when they are stored, and rotated back and xor'ed again when they are retrieved from memory, which recovers the original pointer.

There is no particular significance to the number 0x30; it just happens to be the offset where that value is stored. You can see in the inline assembly that this number comes from offsetof (tcbhead_t, pointer_guard); so the storage at the fs base address is laid out as a tcbhead_t struct, and given the other members that it contains, the pointer_guard member has ended up at offset 0x30. So looking at the name pointer_guard for the member is more informative than its numerical offset.

You can try disabling hardware acceleration using app.disableHardwareAcceleration() (See the docs). I don't think this is a fix though, it just makes the message go away for me.

main.js

Since stdout is line buffered, putchar doesn't write to the terminal directly; it puts the character into a buffer, which is flushed when a newline is encountered. And the buffer for stdout happens to be located on the heap following your heap_book allocation.

So at some point in your copy, you putchar all the characters of your secretinfo method. They are now in the output buffer. A little later, heap_book[i] is within the stdout buffer itself, so you encounter the copy of secretinfo that is there. When you putchar it, you effectively create another copy a little further along in the buffer, and the process repeats.

You can verify this in your debugger. The address of the stdout buffer, on glibc, can be found with p stdout->_IO_buf_base. In my test it's exactly 160 bytes past heap_book.